thatnerd

Back up

I think we've collectively DDoSed it. I'm getting a 504 timeout.

The author [page](https://github.com/sterlingcrispin) is there on github, but I can't even find his full list of his repos to confirm it's still there (I also get a 504 on that).

I ran into them at Burning \Man in the early aughts. They're awesome!

Regardless of how fast we use up non-renewable resources, they're all going to be gone at some point. Copper, lithium, and tin are going to be gone. Humanity will need to live off of what we can forage or grow.

Also, the rarity of farming in the animal kingdom makes me worried about the sustainability even of multi-species domestication. A few ants cultivate trees or fungi or aphids, but they seem to specialize in just domesticating just one species at a time. This is telling us something important: I suspect domesticating too many species leads to vulnerabilities to so many parasites/bacteria/viruses/pests that pestilence and famine risk will eventually outweigh any benefits of domestication. If they didn't, ants would be farming lots of species!

In the real long term, then, humans will get one (or zero) domesticated species, and maybe some electricity if we can make self-sustaining solar power operations using common elements like aluminum and silicon from dirt, or sodium, chlorine, oxygen, and hydrogen from water, and that'll be it for technology, Everything else will be foraged animals and plants, in an ecosystem that keeps our population in check through predation.

As for the transition, it's going to suck. And I don't trust any governing body to "ramp down" the population smoothly without committing some major atrocities.

Yes: our politicization of science and theirs will take a somewhat different form. And yes: if our politicization of science is is less complete, or briefer, it may not set us back as much... but RFK Jr. is out there purging scientists from the federal government if they disagree with his preferred theories about vaccines. This is not going to advance medical science in the US.

The comparison to the present-day US isn't hyperbole. It's not a perfect match but there are parallels.

The article mentions Lysenko, a Soviet "biologist" who set back Soviet biology for a generation. He believed, for example, that plants in the USSR would not compete with each other for resources the way they did in capitalist societies, but would instead share resources. He asserted that crops could therefore be planted closer together in the USSR, yielding more food per acre. Evidence to the contrary was suppressed: Lysenko had Stalin's ear and a zealot's confidence. The rest of the field was either purged or fell in line. Scientists lost their jobs or got sent to Siberia.

The comparison to the present-day US isn't perfect, but it also isn't hyperbole. While scientists in the US mostly aren't in the same type of danger of arrest for speaking out (assuming ICE doesn't start targeting political opponents), but we're looking at a similar era in the US in terms of making theories and data fit ideology. RFK, Jr. has his preferred biological theories about vaccines, autism, and disease. Government scientists are at risk of losing their jobs and their financial security if they reference (or publish) findings that Kennedy objects to. Universities are still (as far as I can tell) safe for natural scientists because the first wave of the crackdown is focused on the humanities and social sciences, so this purge of scientists is limited to federal government employees, but the effect is real, and it isn't a stretch to assume that if the government finds success in the current purge, it will go looking further afield.

The human impact is significant for those affected, but the article is right to point out that this purge of scientists from the government for ideological goals will have a broader impact for society: it will set back American science.

Kennedy doesn't even have to be wrong on the facts for the culture he's creating to be toxic for federal science in his department and beyond. Just the politicization of science pushes our country towards being a scientific backwater.

https://www.wired.com/story/worldcoin-sam-altman-orb/

For a casino? In practice, yes, fair games are perfectly consistent with greedily skimming a game, and fair games draw gamblers.

That said, when organized crime gets involved, somebody always thinks "if I rig this, I'll do EVEN BETTER!" Maybe they're a corrupt employee skimming from the house, maybe they're a loyal employee skimming for the house, but unless you have something like the Nevada Gaming Control Board forcing fairness on them, you basically never get it. At least, from what I've read on the subject. Source: I've read some books on card counting & otherwise beating the odds in casinos, and this my vague memory.

And it's ironic that the house wants to rig games, because a biased game means a mathematically savvy individual can go in and calculate how results differ from "fair" games, and can then skim some profits for themselves if the bias is larger than the house advantage.

I think that's an invalid hypothesis here, not just an unlikely one, because that's not my understanding of how LLMs work.

I believe you're suggesting (correctly) that a prediction algorithm trained on a data set where women outperform men with equal resumes would have a bias that would at least be valid when applied to its training data, and possibly (if it's representative data) for other data sets. That's correct for inference models, but not LLMs.

An LLM is a "choose the next word" algorithm trained on (basically) the sum of everything humans have written (including Q&A text), with weights chosen to make it sound credible and personable to some group of decision makers. It's not trained to predict anything except the next word.

Here's (I think) a more reasonable version of your hypothesis for how this bias could have come to be:

If the weight-adjusted training data tended to mention male-coded names fewer times than female-coded names, that could cause the model to bring up the female-coded names in its responses more often.

Yes, the pricing is what I'm curious about. If they were saying "we can make this for half the price of steel," then the world steel market becomes their growth target.

If they can't get cheaper than steel, if they can't compete with steel, they'll probably never be more than a niche product.

Here's a thought: why don't YOU shut up? /s

Sorry! That was mean, but I hope it came across as funny.

In all seriousness, I like the question, and your implication is intuitive: if we (as individuals) talk to machines rudely, it's likely to (at minimum) lead us to be ruder to other humans, if only by habit. And if they're expecting more politeness than we're showing, they may infer the intent to be rude, and react accordingly. Those who are rude would end up being worse off.

That said, it's the Fallacy of Composition to assume that if everyone gets ruder the collective effect would be the same as the individual effect. We have different requirements for what counts as "polite" in different cultures but everyone seems to get along pretty well. Maybe societies can all get ruder (and just get along worse with each other) but also maybe they can't.

I tried looking in the literature but this book implies we don't even know how to measure politeness differences between languages: https://books.google.com/books?hl=en&lr=&id=MPeieAeP1DQC&oi=...

There are even theories that politesse can lead to aggression: https://www.jstor.org/stable/2695863

Deborah Tannen (the linguist) has found many examples where different politeness expectations (particularly across the cultural divide that aligns with gender) can lead to conflict, but it always seems to involve misunderstandings due to expectations: https://books.google.com/books?hl=en&lr=&id=YJ-wDp7CJYAC&oi=...

So yeah, bad outcomes feel intuitive but I don't think linguistics or sociology has a theory of what happens if a group collectively gets less polite.

Sam Altman doesn't want you to say "please" to ChatGPT.

https://futurism.com/altman-please-thanks-chatgpt

You can't express mass in volts. A volt is energy per unit of charge. To get energy, you need to multiply by a charge.

One Joule of energy is what you get when you move one Coulomb of charge across a 1V potential.

One electronVolt (eV) is the energy you get from moving one electron's worth of charge across 1 volt of potential.

It's an accident of what we chose to be a Joule of energy and what we chose to be a Coulomb of charge, so there should be no expectation that this would turn out to be the mass of an electron (when divided by the square of the speed of light, which is unstated because everyone knows E = mc^2).

I think it's an upper limit of around 0.2 eV^2, so √(0.2 eV^2) = 0.45 eV.

I'd like to see an experiment that's sensitive enough to eliminate the m=0 hypothesis. Their prediction of m^2 = -0.14eV^2 is a bit disconcerting.

What is the stochastic part? It looks like the predator/prey behavior is deterministic.

I'm guessing it's somewhat similar to the foxes/rabbits work you were doing a few months ago? https://github.com/kylebebak/foxes_and_rabbits/blob/main/fox...

I think the demographic transition overshoot (https://en.wikipedia.org/wiki/Demographic_transition#Stage_f...) we see is unexpected though.

Yeah the demographic transition is something nobody predicted (afaik). On the other hand, LTG (https://en.wikipedia.org/wiki/The_Limits_to_Growth) is a neo-Malthusian prediction that seems to match early data, and a surprising number of people revisit it and find its conclusions seem to hold. We'll be finding out around 2040, give or take. ¯\_(ツ)_/¯

Feedback: The units on square feet per acre are inverted. Should be acres per square foot.

I recommend having the solution check unit cancellation to confirm that it works.

I'd also like to see a choice between multiplying/dividing in the equation.

We're already engineering ecosystems in a thoughtless way. I find it hard to believe that messing with mosquito genetics is going to make the world significantly more risky than it already is for humanity.

This article is too strong on vibes and lacks data beyond "there's a lot more money in private equity than there used to be."

Also, private equity includes a broad range of investors, from angels that fund unlikely-to-ever-be-profitable startups to leveraged buyout firms that take large companies private because they think they can do better than current management, to vulture funds that snap up bankrupt companies.

I'm not even clear what the author is talking about here.

Location: New York/New Jersey

Remote: Yes

Willing to relocate: No

Technologies: Python, shell script, SQL, Relational and NoSQL Databases

Resume/CV: https://www.linkedin.com/in/williamdcross/

Email: thatnerd (at) gmail.com

William Cross, Ph.D.

Technical curriculum developer with 10 years of experience (primarily at database startups). Educational background is in experimental physics.

Useful for bored apes this week

That's always been my take too.

Like, ok, we get testable answers and they match experiments but also this is _so_ hacky and I can't shake the feeling that one day someone will come along and show that there's some reason why these bad assumptions work out fine. You know, like how "to find the Schwarzschild radius for a black hole of known mass, calculate the radius at which the escape velocity is equal to the speed of light" gives the correct answer even though the theory implied by this method is naive and wrong.

Thank you, we didn't get that resolved beforehand. Should be fixed now.